namic values were obtained from radial pulse waveforms using theSpyghmoCor device (AtCor Medical Inc. USA), incorporating soft-ware that analyzes direct applanation tonometry. Aortic waveformswere constructed and arterial compliance surrogates, AP and AIx-75(standardized to pulse of 75), were derived. An operator index of � 80assured consistency. After delivery the subject’s chart was reviewed toobtain final diagnosis. The AP and Aix-75 values were compared tofinal diagnosis. Statistical analyses were performed using ANOVA andTukey multiple comparison tests and ROC curves were created in SAS9.2.RESULTS: 151 women were recruited with an average gestational age of35.6 weeks. 76.2% of the cohort had comorbidities: GHTN(33.9%),CHTN(41.7%), asthma(7.8%), diabetes(5.2%), other (11.3%). Readingswith operator index �80 were eliminated; 136 were analyzed. 44% of thecohort was diagnosed with preeclampsia (mild�63.6%, severe�12.1%,superimposed�24.2%); these 59 AP and 54 Aix-75 readings were com-pared to the final diagnosis. AP and AIx-75 readings were significantlycorrelated to the diagnosis of preeclampsia (p�.005 and p�.004 respec-tively). AP readings were significant in diagnosing severe preeclampsia(p�0.03) and in distinguishing preeclampsia from chronic hypertension(p�0.03). ROC curves were constructed; an AP value of 9.5 and anAIx-75 value of 27.1 were 80% sensitive in predicting the diagnosis ofpreeclampsia, but specificity was only 40%; area under AP and AIx-75curves were 0.62, and 0.65, respectively.CONCLUSION: Applanation tonometry may be a useful adjunct in mak-ing the diagnosis of preeclampsia, particularly in individuals in whomthe diagnosis is not clear. However test characteristics of the AP andAIx-75 are not as high as desired.

646 The recurrence risk of preeclampsia in subsequentpregnancies in northern Tanzania: a registry–basedprospective cohort studyMichael Mahande2, Anne Dalveit2, Blandina Mmbaga2, JosephObure3, Rachel Manongi1, Rolv Lie2

1Kilimanjaro Christian Medical College, Community Health, Moshi, UnitedRepublic of Tanzania, 2University of Bergen, Public Health and Primaryhealth care, Bergen, Norway, 3Kilimanjaro Christian Medical C, Obstetricsand Gynaecology, Moshi, United Republic of TanzaniaOBJECTIVE: To compare the risk of preeclampsia in subsequent preg-nancies between women who had preeclampsia, and women who didnot have preeclampsia in a previous pregnancySTUDY DESIGN: Prospective cohort studyRESULTS: A total of 3,909 women who were followed returned for theirsubsequent pregnancies. The overall risk of preeclampsia in the firstpregnancy was 3.5% compared with 3.2% in subsequent pregnancies.Preeclampsia reoccurred in 24.6% of women in subsequent preg-nancy. The recurrence risk of preeclampsia was 9.2 (95% CI: 6.4 –13.2) times higher in women with prior preeclampsia. Maternal andfetal conditions in the first pregnancy that increase women’s likeli-hood risk of preeclampsia in the subsequent pregnancy were pretermbirth (RR� 3.1; 95% CI: 2.1– 4.7), perinatal death (RR� 3.9; 95% CI:2.9 –5.9), low birthweight (RR� 3.1; 95% CI: 2.1– 4.5), chronic hy-pertension (RR� 8.9; 95% CI: 5.7–13.8), and gestational hyperten-sion (RR� 9.8; 95% CI: 4.9 –19.1). Continuation rate to the nextpregnancy decreased after preeclamptic pregnancy.CONCLUSION: The recurrence risk of preeclampsia was higher as thatreported in most studies in the high income countries. Prior pre-eclampsia is the strongest risk factor for a recurrent preeclampsia.Early identification of women at risk during prenatal care, especiallythose with history of preeclampsia will help clinicians in designingbetter management and prevention of maternal and fetal morbidityand mortality related to eclampsia.

647 MicroRNA 210 levels in the first and second trimester ofpregnancy are accurate predictors of pregnancy relatedhypertensionMichal Elovitz1, Lauren Anton1, Jamie Bastek1, Nadav Schwartz1

1University of Pennsylvania, Maternal and Child Health Research Program;Dept OBGYN, Philadelphia, PAOBJECTIVE: MicroRNAs (miRNAs) are highly conserved single-stranded non-coding RNA molecules that play a critical role in generegulation. MiRNA517a, miR-517c and miR-210 have emerged as po-tential biomarkers for pregnancy related hypertension (PRH). Theobjective of this study was to determine if select miRNAs levels in thefirst trimester could predict PRH and if serial levels (1st and 2ndtrimester) provided better prediction of PRH.STUDY DESIGN: A nested case control was performed from a largerprospective cohort of singleton pregnancies in which maternal bloodwas collected at 11-13 week (TRI-1) and 15-20 weeks (TRI-2). Caseswere defined as women with PRH (n�39), controls were women whodelivered without PRH at term (N� 55). MIR-517a, miR-517c andMiR-210 levels were assessed. Analyses included non-parametric teststo compare medians, MVLR to estimate odds, and ROCs to determineto predictive ability of each miRNA. The difference and ratio betweenTRI-1 and 2 miRNA levels were also assessed and modeled to deter-mine their predictive ability.RESULTS: In TRI-1 samples, only miR-210 levels differed significantlyby case status (Table). TRI1 miR-210 levels were significantly different(�0.001) as was the MI-DIFF (P�0.015). Controlling for race, age,screening BMI, and nulliparity, miR-210 increased the odds of beinga case of 5.7-fold (95% CI: 2.1-15.6 P�0.001). When the miR-210difference was investigated by tertiles, the highest tertile increasedodds of PRH 9.5-fold (95% CI: 2.2-41.3 P�0.003). In the most par-somonius model, ROC demonstrated an AUC of 84. When investi-gating the ratio of miR210 (TRI2/TRI1), for each unit of change, theodds of being a case was 1.4 (5% CI: 1.12-1.64 P�0,001) with an AUCof 90 (Figure).CONCLUSION: TRI1 and TRI1 MiR-210 levels are highly predictive ofPRH months prior to the onset of clinical symptoms. The value ofsuch miRNA biomarkers should be explored in larger cohorts to de-termine their utility for clinical use.

Ratio of TRI2/TRI1 miRNA-210levels in predicting PRH

www.AJOG.org Academic Issues, Antepartum Fetal, Clinical Ob, Fetus, Genetics, Hypertension, Med-Surg-Diseases, Operative Ob, U/S Poster Session IV

Supplement to JANUARY 2013 American Journal of Obstetrics & Gynecology S273

648 The impact of change in pregnancy body mass index ongestational hypertension/preeclampsiaMorgan Swank1, Aaron Caughey2, Christine Farinelli1, ElliottMain3, Kathryn Melsop3, William Gilbert4, Judith Chung1

1University of California, Irvine, Maternal Fetal Medicine, Orange, CA,2Oregon Health & Science University, Maternal Fetal Medicine, Portland,OR, 3California Maternal Quality Care Collaborative, CMQCC, Stanford,CA, 4Sutter Medical Center, Maternal Fetal Medicine, Sacramento, CAOBJECTIVE: To examine the impact of changes in body mass index(BMI) during pregnancy on the incidence of gestational hypertension(GHTN)/preeclampsia (PRE).STUDY DESIGN: This is a retrospective cohort study using linked birthcertificate and discharge diagnosis data (All-California, Rapid-Cycle,Maternal/Infant Database) from the year 2007. Adjusted odds ratios(aOR) with 95% confidence intervals (CI) were calculated for theoutcome of GHTN/PRE, as a function of a categorical change in preg-nancy BMI: BMI loss (BMI change��0.5), no change (�0.5 to 0.5),minimal (0.6 to 5), moderate (5.1 to 10), and excessive (�10). Nochange in BMI served as the reference group. The impact of pregnancychange in BMI was determined for the entire cohort and then strati-fied by prepregnancy WHO BMI category.RESULTS: The study population consisted of 436,414 women with sin-gleton gestations. Overall, women with excessive BMI change had anearly two-fold increased incidence of GHTN/PRE, when comparedto women with no net change in BMI (aOR�1.94; 95% CI�1.72-2.20). By prepregnancy BMI class, the aOR for GHTN/PRE in womenwith excessive BMI change were: normal weight (aOR�2.74, 95% CI,1.87-4.04), overweight (aOR�3.43, 95% CI�2.62-4.94), obese class I(aOR�2.47, 95% CI�1.93-3.18), obese class II (aOR�2.53, 95%CI�1.84-3.48), obese class III (aOR�2.89, 95% CI�2.00-4.19).CONCLUSION: Regardless of prepregnancy BMI category, the aOR ofdeveloping GHTN/PRE was significantly higher in women with BMIchange �10 when compared to women with no or minimal BMIchange. Women who entered pregnancy in the overweight category(BMI�25.0-29.9) were at the greatest risk for developing hyperten-sive disorders with excessive BMI change. This may provide beneficialmaterial for counseling patients of normal and overweight status, whoare typically considered to be of a lower risk strata. Unadjusted inci-dence of gestational hypertension/preeclampsia as a function ofchange in pregnancy body mass index

649 PTX3, sFlt-1 and PlGF levels in mothersand their own newbornPaola Algeri1, Sara Ornaghi1, Davide Bernasconi2, FabrizioCappellini3, Stefano Signorini3, Paolo Brambilla4, GabrieleUrban5, Patrizia Vergani1

1University of Milan-Bicocca, Obstetrics and Gynaecology, Monza, Italy,2University of Milan-Bicocca, Clinical medicine and Prevention, Center ofBiostatistic for Clinical Epidemiology, Monza, Italy, 3University of Milan-Bicocca, Laboratory Medicine, Desio, Italy, 4University of Milan-Bicocca,Experimental Medicine, School of Medicine, Milan, Italy, 5University ofMilan-Bicocca, Obstetrics and Gynaecology, Desio, ItalyOBJECTIVE: An alteration of PTX3, sFlt-1 and PlGF serum levels isdescribed during preeclampsia. Instead, no data about these markersin fetal serum are reported in literature. The aim of our study is tocompare serum maternal and fetal markers’ levels in order to evaluatea potential relation between them.STUDY DESIGN: A case-control study with collection of 74 maternalserum samples and 74 fetal serum samples obtained from umbilicalblood of each own baby. Samples were collected at time of delivery in22 preeclamptic women and their newborns (cases), 23 women andtheir own babies, who formed a gestational age homogeneous groupcompared to cases (including pregnancies complicated by pretermdelivery, IUGR and pPROM), and 29 healthy controls and their new-borns (normal pregnancies delivered by elective cesarean section at �37 weeks). Statistical analysis was performed with Paired Wilcoxontest and Spearman correlation.RESULTS: Maternal and fetal levels of PTX3, sFlt-1 and PlGF werefound to be different (p � 0.0001 for each comparison). For eachmarker, the fetal level median was lower than the maternal one (PTX3:maternal median 1.78 ng/ml vs fetal median 1.11; sFlt-1: maternalmedian 5001.5 pg/ml vs fetal one 242.6; PlGF: maternal median 94.3pg/ml vs fetal one 11.2).Table 1 shows Spearman correlation results:an independent trend for PTX3 in maternal versus fetal serum (p �0.14) was found, whereas maternal and fetal levels of sFlt-1 and PlGFpresented a directly proportional relationship.CONCLUSION: Our results support a possible independent productionof PTX3 in the fetus, whereas a trans-placental passage from mother tofetus is more likely to be the mechanism underlying fetal sFlt-1 and

Demographics and miRNA levels betweencases and controls

miRNA levels are determined by QPCR and normalized to U6.

Poster Session IV Academic Issues, Antepartum Fetal, Clinical Ob, Fetus, Genetics, Hypertension, Med-Surg-Diseases, Operative Ob, U/S www.AJOG.org

S274 American Journal of Obstetrics & Gynecology Supplement to JANUARY 2013